$\texttt{GENGARS}$: Accurate non-Gaussian initial conditions with arbitrary bispectrum for N-body simulations

TL;DR

GENGARS is a new framework for generating accurate non-Gaussian initial conditions with arbitrary bispectra for N-body simulations, improving precision over existing methods and enabling detailed studies of primordial non-Gaussianity.

Contribution

It introduces GENGARS, a novel method for creating non-Gaussian initial conditions with arbitrary bispectra, outperforming previous techniques in accuracy and noise reduction.

Findings

GENGARS achieves higher accuracy than 2LPT-PNG in simulations.

The method effectively suppresses spurious power spectrum contributions.

GENGARS is versatile for different non-Gaussian templates.

Abstract

Primordial non-Gaussianity is predicted by various inflationary models, and N-body simulations are a crucial tool for studying its imprints on large-scale structure. In this work, we present \texttt{GENGARS} ( GEnerator of Non-Gaussian ARbitrary Shapes), a framework for generating accurate non-Gaussian initial conditions for N-body simulations. It builds upon the formulation introduced by Wagner \& Verde (2012), enabling to generate a primordial gravitational potential with a desired separable bispectrum $B_{\Phi}(k_1,k_2,k_3)$. For the local, equilateral and orthogonal non-Gaussian templates, we benchmark our method against the well-established \texttt{2LPT-PNG} code. We show that \texttt{GENGARS} achieves improved accuracy and lower noise by suppressing spurious contributions to the primordial power spectrum. This paper aims at presenting the method, quantifying its performance and illustrating the benefits and applicable use cases over existing approaches.